Challenges and Solutions for the Next Generation of Vaccines: Jonathan Liu Luis Maranga Sachin Mani Richard Schwartz

Transcription

1 Challenges and Solutions for the Next Generation of Vaccines: Development of fcell llculture-based dlive Attenuated t Influenza Vaccine VACCINE TECHNOLOGY II June 1, 2008 Albufeira (near Faro), Algarve, Portugal Jonathan Liu Luis Maranga Sachin Mani Richard Schwartz

2 Influenza A Major health problem worldwide, each year 5-15% of the population are affected with upper respiratory tract infections 3-5 million cases of severe illness 250, , deaths Pandemic and seasonal flu Vaccination - the principal measure to prevent the disease and reduce the impact Caused by influenza virus 8 segmented negative sense RNA genome encode 11 proteins including 2 surface antigens: HA and NA Antigenic differences in HA and NA determine virus type (influenza A viruses) and lineage (influenza B viruses) / /f /f / / Kaiser et al., Science, Vol 312, p

3 FluMist FluMist (Influenza Virus Vaccine Live, Intranasal) Cold adapted, live attenuated vaccine Innovative technology (nasal administration) Antigen sparing (high yield) Durable mucosal and systemic immunity High efficacy* *Belches et al., NEJM, 356 (7), p

4 Manufacture of FluMist Influenza Vaccines Wt Selection (1/2 8 or 1/256) Cloning 6:2 reassortant vaccine HA NA Transfection PB1 PB2 PA NP M NS HA NA MDV WT Donor strain ca A/Ann Arbor/6/60 or ca B/Ann Arbor/1/66 (MedImmune) Manufacture Plasmids Containing genes for vrna Blending of three strains (ca, ts, att) Eggs FluMist (Influenza Virus Vaccine Live, Intranasal) 4

5 Old and New Influenza Vaccine Production Technologies Challenges with egg production platform Egg stock vulnerability Production capacity limited Less defined biological starting material and significant operator intervention Egg allergies Advantages of cell culture-based production platform Susceptibility to a broad spectrum of influenza virus strains Better defined production substrate More controlled manufacturing process sequestered from the environment Surge capacity flexibility Rapid scale-up 5

6 Production Cell: Madin-Darby Canine Kidney (MDCK) cells 6

7 Selection of Host Cell for Cell Culture Flu Vaccine Production Thirteen cell lines (9 mammalian and 4 avian host cells) were tested MRC-5 5, WI-38 38, VERO, FRhL-2, 293, NIH 3T3, CHO, MDCK and other human cell lines CEF, CEK, DF-1 and avian embryonic stem cell line Only Vero and MDCK cells produced viruses >6.0 log 10 TCID 50 /ml for FluMist strains Only MDCK cells produced viruses with titer >7.0 log 10 TCID 50 /ml for all types and families of seasonal strains tested 7

8 Challenges Related with Use of MDCK cells for Vaccine Production Original line of MDCK cells was non tumorigenic Some MDCK derivatives have been found to be highly tumorigenic Highly tumorigenic cell substrates have never been used to manufacture ac u viral vaccines Highly tumorigenic cell substrates pose significant regulatory challenges Krause, VRBPAC 2005 Product Safety and Regulatory Concerns! 8

9 MedImmune s Approach to Minimize Risk of MDCK Cells Produced cell bank with low tumorigenic potential from biologically cloned cells Extensive testing strategy developed Process developed to deliver vaccine with the following characteristics Sterile product Acellular Reduction of host cell DNA quantity Reduction of host cell DNA size Minimal exposure to animal derived components (ADCs) with guidance from CBER for: Adventitious agents Tumorigenicity of live, intact cells Oncogenicity of host cell DNA and lysate 9

10 Produced Low Tumorigenic Cell Bank - Development of Serum-free MDCK Cells Numerous in-house SFM formulations developed Cells evaluated for several primary factors Maintenance of cell line growth Potency (influenza virus yield) Karyology Tumorigenicity in athymic nude mouse model 10

11 Tumorigenicity of 10 7 Uncloned MDCK Cells in Adult Nude Mice Treatment Group (n=10) Animals with Tumors Tumor Formation Rate Note MDCK Cells in Serum medium 0 0% MDCK Cells in SFM A 6 60% tumors at injection site MDCK Cells in SFM B 0 0% Negative Cell Control 0 0% Positive Cell Control % tumors at injection site 11

12 Produced Low Tumorigenic Cell Bank - MDCK Cells Cloning Began after development of SFM that maintained cell karyology and low potential tumorigenic nature of the ATCC MDCK cells Initiated from a new vial of ATCC MDCK cells Performed in serum-containing i media Two rounds of limiting dilution cloning completed Clones initially iti selected based on productivity it 12

13 Initial Cloning of MDCK Cells Number of Clones < Range of Virus Titer (log FFU/mL) Range < < 7.6 to 8.8 Clone No Percentage 82.6% 8.6% 3.1% 2.9% 1.3% 0.9% 0.5% 0.1% 0.2% 100% 13

14 Not All High-Producing Clones in SFM Maintained Cell Growth 0^5 cells/ml) Cell Density (x1 Viable MDCK Subclone A Cell Passage Number Gradually reduced growth rate in SFM No longer growing after 11 passages Eliminated from clone selection 14

15 Extended MDCK Cell Growth in SFM 2.50E E+07 Subclone C Subclone B 1.50E E+07 (D4) 5.00E+06 (D3) 0.00E+00 P1 (D4) P2 (D4) P3 (D3) P4 (D3) P5 (D4) P6 (D4) P7 (D4) P8 (D4) P9 (D3) P10 (D3) P11 (D4) P12(D4) P13 (D4) P14 (D3) P15 (D3) P16 (D4) P17 (D3) P18 (D4) P19 (D3) P20 (D4) P21 (D3) P22 (D3) P23 (D4) P24 (D3) P25 (D3) Number of Passage in SFM Total C ells per T75

16 Tumorigenicity and Oncogenicity of Cloned MDCK Cells Study Test Sample Animal with Tumors/ Total Animals Location of Tumors Tumorigenicity DPBS 2/33 lymph node, spleen, liver adult nude mice Hela cells 38/41 site of injection 10 5 MDCK cells 1/44* spleen, liver, lung 10 1,10 3,10 7 MDCK cells 0/132 Tumorigenicity Hela cells 44/44 site of injection NB nude mice 10 1, 10 3, 10 5, 10 7 MDCK cells 0/176 * tumors confirmed by histology examination, not sharing MDCK cell morphology, not located at SOI and not related with MDCK cells as verified by immunohistochemistry analysis Test Samples Newborn Animals non-injected Saline MDCK cell lysate MDCK cell DNA mice 1/25 0/45 0/45 na n.a. MDCK Cell Lysate hamsters 0/25 0/45 0/45 n.a. Oncogenicity rats 0/25 0/45 1/45* n.a. mice 0/25 0/45 n.a. 1/45* MDCK Cell DNA hamsters 0/25 0/45 n.a. 0/45 Oncogenicity rats 0/25 1/45 na n.a. 0/45 16

17 Production Platform: A Flu Vaccine Manufacture Process without Extensive Process Development 17

18 Challenges Related with Annual Flu Vaccine Production Surveillance Selection Reassortment Manufacturing Distribution Immunization JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Virus Strain A (H3N2) A/Guizhou/54/1989 A/Beijing/353/1989 A/Beijing32/1992 A/Shandong/9/1993 A/Johannesburg/33/1994 A/Wuhan/359/1995 A/Sydney/5/1997 A/Moscow/10/1999 A/Fujian/411/2002 A/California/7/2004 A/Wisconsin/67/2005 A/Brisbane/10/2007 like A (H1N1) A/Singapore/6/1986 A/Bayern/7/1995 A/Beijing/262/1995 A/New Caledonia/20/1999 A/Solomon Islands/3/2006 A/Brisbane/59/2007 like B B/Yamagata/16/1988 B/Panama/45/1990 B/Beijing/184/1993 B/Sichuan/379/1999 B/Hong Kong/330/2001 B/Shanghai/361/2002 B/Malaysia/2506/2004 B/Florida/4/2006 Frequent vaccine strains and manufacture process changes Short (often 1-3 weeks) Process Development Time 18

19 MedImmune s Approach to Address Process Development Challenges Understand process parameters to all strains Develop a true platform process capable of production of all strains Conduct DOE and reduce pre-production PD time to approx. 2-4 weeks 19

20 Time of Infection and Virus Yield FFU/cell FFU/cell Time of harvest: 2 dpi Virus: A/Wisconsin/67/05 irus titer (Log 10 FFU/ml) Vi FFU/cell Viable cell density cells/ml) Viable cell density (1 Data collected using shake flasks Virus titer estimated using FFA and viable cell density estimated using nuclei counts Error bars correspond to assay replicates and biological replicates Time of infection (days post seeding) Peak titer increases with the time of infection (e.g., 4dps vs 3dps) irrespective of the virus input at the time of infection Trends are similar among different sub-types 20

21 Input Virus and Virus Yield Virus: A/Wyoming/03/ Data collected using shake flasks Virus titer (Log 10 FF FU/ml) FFU/cell FFU/cell FFU/cell FFU/cell FFU/cell Virus titer estimated using FFA and viable cell density estimated using nuclei counts Error bars correspond to assay replicates Hours post infection (hpi) Lower virus input improves virus titer 21

22 Spread of Virus Productivity Using Phase I Platform Process B strains H1N1 H3N2 Pandemic H5N1 H7N3 H9N B/Florida/7/04 /04 B/Yamanashi/166/98 /98 B/Victoria/504/ B/Ann Arbor/1/66 /66 B/Malaysia/2506/04 /04 A/Texas/36/91 A/Solomon Islands/3/06 A/Hong Kong/2652/06 A/Beijing/262/95 A/New Caledonia/20/99 /99 A/California/07/ A/Nepal/921/06 A/Wyoming/03/ A/Wuhan/395/95 A/Panama/2007/99 A/Wisconsin/67/05 A/Hong Kong/213/ A/Hong Kong/491 H5_486 N1/1997 A/Vietnam/1203/ A/British Columbia/06/ A/Hong Kong/G9/ Cold adapted vaccine strains 22 Virus titer (Log1 FFU/ml) 10

23 Risks Posed by Virus Productivity Variation ~ 60 doses/ml Doses/ml increases B/Florida/7/ /7/04 B/Yamanashi/166/ 66/98 B/Victoria/504/ B/Ann Arbor/1/ /1/66 B/Malaysia/2506/ 06/04 A/Texas/36/ 6/91 A/Solomon Islands/3/ 3/06 A/Hong Kong/2652/ 2/06 A/Beijing/262/ 2/95 A/New Caledonia/20/ 20/99 A/California/07/ A/Nepal/921/ 1/06 A/Wyoming/03/ A/Wuhan/395/ 5/95 A/Panama/2007/ 7/99 A/Wisconsin/67/ 7/05 A/Hong Kong/213/ A/Hong Kong/491 H5_486 N1/ A/Vietnam/1203/ A/British Columbia/06/ A/Hong Kong/G9/ 9/ ~ 1 dose/ml Cold adapted vaccine strains 23 Virus titer (Log10 FFU/ml)

24 Improvement in Process Performance and Robustness Pre-optimization Post-optimization B strains H1N1 H3N2 Pandemic B strains H1N1 H3N2 Pandemic B/Florida/7/04 B/Yamanashi/166/98 B/Victoria/504/2000 B/Ann Arbor/1/66 B/Malaysia/2506/04 A/Texas/36/91 A/Solomon Islands/3/06 A/Hong Kong/2652/06 A/Beijing/262/95 A/New Caledonia/20/99 A/California/07/2004 A/Nepal/921/06 A/Wyoming/03/2003 A/Wuhan/395/95 A/Panama/2007/99 A/Wisconsin/67/05 A/Hong Kong/213/2003 A/Hong Kong/491 H5_486 N1/1997 A/Vietnam/1203/2004 A/B /British Columbia/06/2004 A/Hong Kong/G9/97 B/Florida/7/04 B/Yamanashi/166/98 B/Victoria/504/2000 B/Ann Arbor/1/66 B/Malaysia/2506/04 A/Texas/36/91 A/Solomon Islands/3/06 A/Hong Kong/2652/06 A/Beijing/262/95 A/New Caledonia/20/99 A/California/07/2004 A/Nepal/921/06 A/Wyoming/03/2003 A/Wuhan/395/95 A/Panama/2007/99 A/Wisconsin/67/05 A/Hong Kong/213/2003 A/Hong Kong/491 H5_486 N1/1997 A/Vietnam/1203/2004 A/Br British Columbia/06/2004 A/Hong Kong/G9/ Cold adapted vaccine strains 6.9 Cold adapted vaccine strains Average productivity improvement = 0.5 log 10 FFU/ml Lowest virus titer observed 7.9 log 10 FFU/ml (up from 7.1 log 10 FFU/ml) Reduced variability in yield, increasing process robustness 24 Virus titer FFU/ml) (Log10

25 Fully Disposable Small Scale Platform Process T75 T225 RB RB 50L Single Use Bioreactor Cell inoculum expansion Virus production days Fully disposable process implemented in GMP Pilot Plant No need for cleaning/validation with disposable culture vessels Shortened timeline for implementation in clinical production Quick turned-around between batches (in a few hours), making possible to re-start production very rapidly. 25

26 Projected Large-scale Manufacturing Process Vial Thaw Cell Factories 50 L Bioreactor 75 cm 2 Flask 225 cm 2 Flasks Cell expansion 500 L Bioreactor Production Bioreactor ~ 2000 L Virus production 26

27 Theoretical Bulk Vaccine Dose Output Number of Trivalent Doses (x10e+6) Purification Yield (%) Harvest Titer (FFU/mL) Trivalent doses for 2x2000L bioreactors, 6 month campaign, 7.0 log 10 FFU dose 27

28 Summary and Outlook With a proprietary p serum-free medium and through biological cloning, MedImmune has prepared a MDCK cell bank that is shown to be Low in tumorigenicity and oncogenicity High in virus productivity for both pandemic and seasonal influenza vaccine strains By identifying and optimizing critical process parameters MedImmune has developed a flu vaccine platform manufacture process that is highly productive for seasonal (H1N1, H3N2 and B) and pandemic (H5N1, H7N3, H9N2) virus strains with yield >8 log 10 FFU/mL Used to complete Phase I GMP clinical production campaigns successfully in spite of a last minute change in H1N1 virus Phase II Process Development is on-going with focuses on process scale up, robustness and shorter development time 28

29 Posters #27 and #28 Poster #27. Development of a Cell Culture Production Platform for Cold-Adapted Live Attenuated Influenza Vaccine (CAIV) Strains of FluMist : Effects and Interactions of fmedium Components, Trypsin, and dinfluenza Virus Family/Type in Process Productivity Poster #28. Development of a Cell Culture Production Platform for Cold-Adapted live attenuated Influenza Vaccine (CAIV) strains of FluMist : Accelerated Development of a Fully Disposable Phase I Clinical Manufacturing Process 29